Minutes of ATF2 weekly meeting, 6 July, 2011

July 6 15:00 - 16:30 (JST time), Webex and Danwa-Shitsu 2, 1 Go-kan, KEK

The new ATF wiki page (http://atf.kek.jp/twiki/bin/view) can be reached here, where the collaboration-page requests your login or registration.

Possibility to detect ground motion (GM) and the ground motion feed-forward, Y. Renier

A GM model was made with three waves at 0.2, 2.9 and 10.4Hz by fitting the GM which has been measured on Friday 12/12/08 at 3pm ( during shift) by B.Bolzon. The model can calculate vertical and horizontal position reproducing the spatial and frequency properties measured at ATF2.

This study is based on simulation with this GM model, where the beam line elements (i.e. magnets etc.) are displaced with respect to the first element by the GM model. The GM along the beam line is measured by a number of Guralp seismometers with the transfer function (TF). So, the GM is corrected by the TF. On-line, only the real part can be corrected ( feed-forward implementation). The tracking code is PLACET. Also, incoming beam jitter ( 6Hz, 100 pulses), BPM noise and sensor TF are included. In this study, all sextupoles are turned off.

First, sensitive elements are identified by making matrices of the GM effect in a plane of elements and the BPMs. The high sensitive elements are selected to have GM sensors. First and last elements have the sensors. Since the number of sensors are limited to 15, 20 and 30, displacements due to the GM of other elements are estimated by linear interpolation.

Next, the GM induced beam displacements are estimated by the 15, 20 or 30 sensors and the estimated ones are subtracted from BPM readings. The 15 sensors are not enough for the estimation, especially in horizontal direction. The 30 sensors are the most preferable for the estimation.

Actually, the incoming jitter exceeds the GM effect by about 100 times. The simulation shows very good removal/correction of the incoming jitter, where the incoming x, x', y, y' and ΔE/E are reconstructed in pulse by pulse and propagated along the beam line. The 20 and 30 sensors have good correlation between the displacements corrected jitter and the GM effect, especially in vertical direction.

Proposed feed-forward aims to correct the GM effects in next pulse with the GM measurements on quadrupoles between the pulses, e.g. 6Hz in this simulation. The principle would be demonstrated at ATF2 for the BPM systems with 100nm resolution and um level reconstruction which have been achieved, if there are sensors and fast-enough correctors.

Open questions are listed below.

Issues to be investigated are also listed below.
Q Glen : In subtracting the incoming jitter, what level of jitters you assumed ?
A : We used values measured in last December, i.e. x, x', y, y' and ΔE/E, by using Gaussian distributions.
Q : What are sources of the errors ?
A : They are reconstructed and incoming errors .
Q : In the feedforward to the GM, what kind of correction do you use? Is it one phase ?
A : It depends on the correctors.
C Glen : You must use the movers at the FF. Also, the FONT at the upstream can correct the incoming jitter.
Q : How many sensors are needed ?
A : The number is 30, which does not depend on the repetition rate.
Q : What is the actual sensor with the Guralp TF ?
A Andrea : It is the CMG 40T. There is CMG 6T in Al casing, so it is light with similar TF.
Q/C Stewart : It seems that the coherence is just an edge at 1.5Hz, while it is good at 6Hz ?
C Andrea : 20 sensors would be enough for the coherence in 4m.
Q Frank : Why the sextupoles are off ?
A : If they are on, non-linear parts appear. I tried it, but the (present) simulation did not work.
C Glen : It should work even if sextupoles are ON.

Nanometer resolution study of the KEK IPBPM, Y. Kim

YoungIm described the electronic scheme, the support structure with measurement of vibration and the installation plan for beam test of 2 KEK High Q IPBPM-blocks at the upstream test area.

The electronics block diagram was shown, where all the signals of reference cavities and IPBPM-cavities in the c-band are down-converted to 714MHz, then the IPBPM converted signals are detected with respect to the reference ones ( outputs from the limiter) . The detected signals with 2 phases, i.e. I and Q, are digitized by the 14bit ADC. The reference signals are also digitized in the ADC for the beam intensity. The converter and combiner of the IPBPM signals are located close to the IPBPMs, while the detector and ADC are put in the Eel's bedroom (DAQ room).

The electronics noise determining the minimum position resolution comes from the first amplifier in the converter. Therefore, the sensitivity/linearity of the converter was measured at the test bench. The output powers (714MHz) were measured to be linear down to the input power of about -120dBm ( 5.712GHz in X, 6.426HHz in Y). Since the 1nm input power is about -100 dBm, the basic performance was verified.

For high resolution study, the mechanical support must have good coherence between the IPBPM blocks. Vibrations were measured with 2 different supports, i.e two post mover and SLAC mover in 2011.2.9 (Wed.) in the tunnel and 2011.3.4 (Fri.) outside of the tunnel, respectively, both with Sugahara-san. Six accelerometers ( sensitive from 0.1 to 100Hz ) were put on the floor, table and IPBPMs. The SLAC mover system was chosen for better performance on the coherence especially in Y.

The IPBPM system will be installed on a granite table at the upstream test area in this summer. There is a switchable beam line so that it will on the beam line only for the beam test.

There are electronics for 3 IPBPM cavities. So, we need one more electronics. Also, we need an adapter plate between the SLAC mover and the granite table for the switchable beam line, mover controller and the Helix cables from inside of tunnel to eel's bed room, too. We will check the electronics and digitizer.

Q : What kind of electronics do you use?
A : It is Honda's one .
Q : You showed 1nm resoution. But, It is the converter resolution.
A : Yes, we will check resolution in the electronics system.
Q : To switch between 2 beam lines, how do you swap them ?
A Terunuma : The whole setup of 2 IPBPMs can be moved out/in by using a big base plate which is the adapter plate.
Q Glen : In general you have two choices for setting SLAC mover with respect to the beam line.
A : I will consider.
C : It was the same as the QEA magnets in Honda's experiment.
Q Stewart : You can reuse the Helix cables for s-band BPMs in the tunnel.
C : there are some comments on the cables.
A : Yes. So, cables for movers are only missing at this moment.
C : Honda's detector has outputs with 4 phases instead of 2.
C : We like to see more quantitative comparison between the two post mover and the SLAC mover, such as relative jitter amplitudes between the 2 IPBPM blocks.

IPBSM status, interlock issue of cooling water system, J. Yan and M.Oroku

First, Jacqueline explained the cooling water (air bubble?) issue reported at previous meeting, 22 June, 2011. The laser could not restart two weeks ago after refilling the chilling water. We suspected air bubbles for this cause. However, we could turn on the laser right before the company came to fix it on last Monday. It seems to be normal. The company changed the filter and adjusted the power. We received advice in digital laser control to see what affected interlock which prevented the laser from restarting. Possible causes are (1) the external cooling water pressure is low, (2) internal cooling water is too close down to limit and dirty, (3) flash ramp degradation and (4) HV system faults etc. .

We checked purity, i.e. it is OK. The filter works well.

It happened again, when we added the cooling water for shortage, i.e. the laser stopped. The purity seems OK. But, the water temperature was high, 30oC since the box stored water is outside the laser room. If this is the reason, laser should recover by itself. Then, this morning, the laser started. We do not know if the temperature was too high, or air bubbles disturbed flow sensor .

We will store water inside laser room, and refill some time ahead of beam time. However, we need to investigate all possibilities.

Next, Oroku briefly reported that the status monitor of IPBSM will be updated for any shift members before this autumn beam run. There are 86 parameters to be monitored in 2 display panels at both the control and DAQ rooms. The panels show that the green status is OK, while the red one is to ask you for calling an IPBSM member.

Q : Why the purity is not problem ?
A : Because we measured it with 6 to 8 x 105 Ω cm , which are good.
Q : How much is the purity to trigger the interlock ?
A : we asked engineers at the company for these values.

Next meeting is 13th July, 2011.